Characterization of idiopathic premature ovarian failure

Size: px

Start display at page:

Download "Characterization of idiopathic premature ovarian failure"

Darcy Martin
6 years ago
Views:

1 FERTILITY AND STERILITY Copyright 1996 American Society for Reproductive Medicine Printed on acid-free paper in U. S. A. Characterization of idiopathic premature ovarian failure Gerard S. Conway, M.D. * Gregory Kaltsas, M.B. Anita Patel, D.M.U. Melanie C. Davies, M.R.C.O.G. Howard S. Jacobs, M.D. Division of Endocrinology, Department of Medicine, University College London Hospitals, London, United Kingdom Objective: To characterize women with idiopathic premature ovarian failure (POF) by their ovarian ultrasonographic appearances to establish the prevalence of follicular activity and relationship to autoimmunity, estrogen status, and historical background. Design: Retrospective analysis of clinical, endocrine, autoimmune, ultrasonographic, and bone densitometry parameters. Setting: Reproductive Endocrinology Clinics of The Middlesex Hospital, London, United Kingdom. Patients: Data from 135 women with idiopathic POF were analyzed. A reference group of 18 women with normal ovarian function, studied in their follicular phase, was used for comparison of endocrine and ultrasound data. A reference group of 57 women with normal ovarian function was used for comparison of bone densitometry measurements. Main Outcome Measures: Serum E2 concentrations, autoantibody screen, ultrasonographic measures of ovarian volume, uterine cross-sectional area and endometrial thickness and dual roentgenogram bone mineral densitometry of the lumbar spine. Results: The detection of ovaries by ultrasound (in 76%) and follicular activity (in 60% of patients) was associated with higher bone mineral density compared with women in whom ovaries could not be identified. Of 13 patients presenting with primary amenorrhea, ultrasonography identified ovaries in 62% and follicles in 38% whereas 38% had positive autoimmunity. Evidence of autoimmunity was found in 31 % of patients overall and these were indistinguishable from the nonautoimmune remainder in every respect. Conclusions: Ovarian follicular activity, previously considered to be rare, as in the "resistant ovary syndrome," is found in the majority of women with POF using pelvic ultrasonography. Patients presenting with primary amenorrhea have a similar degree of ovarian function, determined by ultrasound, and autoimmunity as those presenting with secondary amenorrhea. The role of autoimmunity in the pathogenesis of POF is not distinguished from nonautoimmune ovarian damage by the measurements made in this study. Fertil Steril 1996;65: Key Words: Premature ovarian failure, resistant ovary syndrome, ultrasonography Cessation of ovarian function before the age of 40 years occurs in 1% of women (1). The diagnosis of hypergonadotrophic amenorrhea in young women always is distressing, particularly when a cause has not been identified. Idiopathic premature ovarian failure (POF) often is presumed to be an autoimmune phenomenon even though in most series autoantibodies are found in only a minority of patients (2-5). In this study, we present an assessment of Received December 27, 1994; revised and accepted August 3, * Reprint requests: Gerard S. Conway M.D., Cobbold Laboratories, The Middlesex Hospital, Mortimer Street, London WIN 8AA, United Kingdom (FAX: ). the utility of routine autoimmune screening III women with POF. Among women with hypergonadotrophic amenorrhea, a subgroup has been defined with "resistant ovary syndrome" in whom the presence of follicles on ovarian biopsy is in contrast to "streak ovaries" with no follicular activity found in POF (6, 7). Despite early reports of pregnancies in women with resistant ovary syndrome (8), the distinction between POF and resistant ovaries has not proved to be of practical value, perhaps because ofthe imprecision of ovarian biopsy. Indeed, ovarian biopsy has proved to be uninformative for either diagnosis or prognosis in that pregnancies have occurred after follicles have failed to be identified on biopsy (3, 9). Conway et al. Characterization of idiopathic POF 337

2 Ovum donation has proved a major advance in the treatment of women with POF. In the future it may be possible, through advances in in vitro ovum maturation, for some women with POF to use their own ova for IVF. We were prompted, therefore, to assess the prevalence of follicular activity in women with POF using ultrasonography in order to define the population for whom future in vitro ovum maturation may be applicable. In this study we present the value of pelvic ultrasound in women with POF, both as a predictor of ovarian activity and to characterize better the "idiopathic" subgroup. We have recorded pelvic ultrasonography with endocrine, autoimmune, and bone mineral density measurements in 135 women with idiopathic POF to determine in particular whether any parameter differentiates between women with and without a clear autoimmune component. MATERIALS AND METHODS Between 1987 and 1993, 323 women presented to the Reproductive Endocrine clinics of The Middlesex Hospital, London, United Kingdom with hypergonadotrophic amenorrhea, defined as serum LH and FSH concentrations> 10 miu/ml (conversion factor to SI unit, 1.00) on two occasions and no menstruation for ~6 months. An underlying cause for POF was found in 131 women; these patients were Turner's Syndrome, 73; chemotherapy, 21; familial POF, 12; pelvic surgery, 7; pelvic irradiation, 6; galactosemia, 6; and 46 XY gonadal dysgenesis, 6. As our purpose was to focus on the 192 women with idiopathic POF, these patients have not been assessed further. To test the utility of ultrasonography, autoimmune screens, and bone mineral density measurements, we restricted this analysis to the 135 subjects who had at least two of these measurements performed in this center. Ultrasonography was missing in 26 (19%), autoimmune screens were missing in 41 (30%), and bone mineral density was missing in 39 (28%) women. The cause of missing data almost entirely was due to nonrepetition of measurements undertaken at other centers and failure of patients to attend for follow-up tests. The clinical history of 135 women with idiopathic POF included 13 with primary amenorrhea and 122 with secondary amenorrhea. The return of spontaneous menstruation was not tested systematically in all patients, however, nine experienced return of menstruation to some degree defined as menstruation on two occasions within any 6 months during follow-up while not on estrogen replacement therapy. Although most patients seeking fertility were referred directly for ovum donation, four pregnancies occurred in the follow-up period of 6 months to 6 years, one spontane- 338 Conway et al. Characterization of idiopathic PDF ously, one with IVF, one after ovulation induction with hmg, and one after receiving 20 mg/d prednisolone for 1 month. The first two pregnancies occurred in women with negative autoimmune screens; the second two had positive autoimmunity. Eighteen volunteers with normal menstrual cycles who were on no medical treatment served as reference group in whom measurements were taken in the follicular phase of the menstrual cycle. Reference bone mineral density measurements were taken from 57 normal women investigated previously (10). Record was made of age of onset, history of autoimmune disorders, height, and weight. Many women in this series were intermittent users of hormone replacement therapy or oral contraceptives after the onset of symptoms and before presentation at this center, making recalled duration of amenorrhea insufficiently reliable for analysis. Blood was drawn on presentation and LH and FSH were measured using polyclonal antisera (W R Butt) and international reference preparations 68/40 and 69/104, respectively, as standards in RIAs (Chelsea Hospital, London, United Kingdom). As some high gonadotropin measurements were not diluted for accurate quantification, LH and FSH measurements were grouped for analysis: LH 10 to 50 miu/ml (49% of cases), LH >50 miu/ml (51%), FSH 10 to 50 miui ml (40%), and FSH >50 miu/ml (60%). Estradiol was measured by direct RIA with solid-phase separation (Baxter Healthcare Ltd, Newbury, United Kingdom) with a sensitivity of 5.4 pg/ml (20 pmol! L), an intra-assay coefficient of variation (CV) of 5%, and an interassay CV of 18%. Autoantibodies against thyroglobulin, thyroid microsomes, parietal cells, intrinsic factor, nuclear antigen, mitochondria, adrenal cortex, steroid-producing cells, and ovary were sought by indirect immunofluorescence as described previously by the Department of Immunology of The Middlesex Hospital (11). For ovarian antibodies, sections of stromal, corpus luteum, and paralutein cells were used with known sera for positive and negative controls. High resolution transabdominal pelvic ultrasound was performed using a Diasonics SPA 1000 sector scanner (Diasonics Incorporated, Milpitas, CA) attached to a 3.5-mHz transducer. Transvaginal ultrasonography also was performed in 18 women whose ovaries were not visualized on transabdominal scans, but in no case did the detection of ovaries improve with the transvaginal approach. Thus, in contrast to other applications of pelvic ultrasonography, such as the polycystic ovary syndrome or ovulation induction, transvaginal scanning was found to be less effective than the transabdominal route in POF because of the more restricted field of imaging, difficulty in defining small ovaries without the con- Fertility and Sterility

The cursors demonstrate a 3-mm follicle in each ovary. trast offered by a distended bladder (Fig. 1), and the failure to obtain uterine dimensions.

3 Figure 1 Examples of transvaginal (left) and transabdominal (right) ultrasound scans in women with POF. Small ovaries are identified more easily using the transabdominal approach because the bowel has been displaced by bladder distention and the resulting added contrast clarifies the ovarian outline. The cursors demonstrate a 3-mm follicle in each ovary. trast offered by a distended bladder (Fig. 1), and the failure to obtain uterine dimensions. Only transabdominal ultrasonography data is presented here. Ovarian volume was calculated as 4/31l-(~ diameter)3, where the diameter was taken as the mean of the height, width, and depth of the ovary. The number of follicles in each ovary was counted and the largest follicle was measured. When two ovaries were identified, a mean ovarian volume and follicle number was calculated. The uterine cross-sectional area was defined as the product of the maximum length and anteroposterior diameter of the uterus measured in the sagittal plane and the maximal endometrial thickness was measured in the plane through the central longitudinal axis of the uterine body. Dual roentgenogram measurements of the bone mineral density of the lumbar spine (L1 to L4) were performed using a Hologic QDR1000 (Hologic, Boston, MA) scanner in 96 subjects. Differences between groups were sought by analysis of variance with Duncan's procedure for multiple comparisons after log transformation of data where appropriate. Associations between groups defined by ultrasonography and autoimmunity were tested with X 2 analysis. Ultrasound and bone mineral density relationships were defined by Pearson's simple correlation coefficient. RESULTS Pelvic ultrasound data were complete in 109 (81 %) patients: of these, ovaries were not visualized in 26 (24%), one ovary was identified in 41 (38%), and both ovaries were identified in 42 (38%). Follicular activity, identified in 65/109 (60%) ultrasound scans overall, was detectable in 31/41 (76%) scans in which one ovary was visualized and in 33/42 (79%) scans that identified two ovaries. mtrasound measurements at presentation were available for seven of nine women who experienced a return of menstruation. The ovaries were identified in each case with clearly identifiable ovarian follicles. None of the women in whom the ovaries were not identified experienced return of menstruation or pregnancy. The clinical data on patients with POF presented in Table 1 confirm estrogen deficiency and reduced bone mineral density, which were more pronounced in those with primary amenorrhea, compared with the reference group. High and low LH groups did not differ in any parameter measured. High and low FSH groups differed only in their mean serum E2 concentrations (16.9 ± 6.9 and 24.2 ± 12.6 pglml, respectively; mean ± SD; P = 0.005). We also have quantified the parallel ultrasound findings of smaller ovaries, uterine dimensions, and endometrial thickness in women with POF. Further comparison of women with primary and secondary amenorrhea revealed no difference in the percentage of women in whom ovaries (62% and 76%, respectively, P > 0.1) or follicles (38% and 60%, respectively, P > 0.1) were identified ultrasonographically or in the percentage with positive autoantibodies (38% and 32%, respectively, P > 0.1). Clinical autoimmune disease was evident in 13 Conway et al. Characterization of idiopathic PDF 339

4 Table 1 Data From Women With POF Compared With Reference Group* Table 2 Comparison of Autoimmune and Nonautoimmune Subgroups* Primary Secondary Reference amenorrhoea amenorrhoea group No. of subjects Age (y) 23 (17 to 29) 28 (17 to 39) 27 (20 to 38) Serum E2 (pg/ml)t 13.7 :+: :+: 9.6=1: 51.5 :+: 23.4 Ovarian volume (cm 3 ) 1.6 :+: 1.1=1: 2.4 :+: 2.0=1: 4.6 :+: 1.4 Endometrial thickness (mm) 2.8 :+: 1.6=1: 4.2 :+: :+: 2.3 Uterine cross-sectional area (cm2) 10.6 :+: :+: 6.3=1: 22.4 :+: 6.3 Bone mineral density (mg/cm2) 0.79 :+: :+: 0.15=1: 1.05 :+: * Values are mean:+: SD; values in parentheses are ranges. t Conversion factor to SI unit, =I: Different from reference group, P < 0.05 Different from secondary amenorrhoea group, P < II Reference group (n = 57). No. of subjects Primary amenorrhea Return of menstruation Age of onset Serum E2 (pg/ml)t Ovarian volume (cm 3 ) Endometrial thickness (mm) Uterine cross-sectional area (cm2) Bone mineral density (mg/cm2) Autoimmune 29 4 (13.8)* 3 (10.3) 26.5 ::': 5.8t 17.6 ::': ::': ::': ::': ::': 0.15 * Values in parentheses are percentages. t Values are means::': SD. t Conversion factor to SI unit, Nonautoimmune 64 7 (10.9) 4 (6.3) 26.9 ::': ::': ::': ::': ::': ::': 0.13 (10%) patients: hypothyroidism (n = 10), Graves disease (n = 1), Addison's disease (n = 2, one of whom was also hypothyroid), and insulin-dependent diabetes mellitus (n = 1). The results of a full autoimmune screen performed in 97/135 patients with idiopathic POF showed that 33 of97 patients (34%) had at least one positive autoantibody test: thyroid microsomal in 24 (25%), thyroglobulin in 19 (20%), gastric parietal cell in 11 (11 % ), pancreatic islet cell in 4 (4%), antinuclear antigen in 4 (4%), intrinsic factor in 3 (3%), steroid-secreting cell in 3 (3%), adrenal gland in 3 (3%), smooth muscle in 2 (2%), and ovary in 1 (1%). The "autoimmune" subgroup was indistinguishable from the 64 (66%) patients with a negative autoimmune screen with regard to age of onset, prevalence of primary amenorrhea, endocrine, ultrasound, and bone mineral density measurements or in the proportion who experienced return of menstruation (Table 2). Three (30%) of 9 patients who experienced return of menstruation had positive autoantibodies. Bone mineral density correlated with ultrasonographic measurements of uterine cross sectional area (r = 0.36, P < 0.001), ovarian volume (r = 0.35, P = 0.001), and endometrial thickness (r = 0.34, P = 0.001) but not with serum E2 concentrations (r = 0.12, P = 0.17). Bone mineral density was lower in patients in whom ovaries were not identified on ultrasound compared with those in whom follicles >4 mm were identified, and, in those with follicles of ::;4 mm, the mean bone mineral density was intermediate (Fig. 2). DISCUSSION Our identification of ovarian follicles by ultrasound in 60% of women with POF compares with 340 Conway et al. Characterization of idiopathic PDF 41 % previously reported in 17 subjects using transvaginal ultrasound (12) and has led us to defer our use of the label "resistant ovaries," defined by evidence offollicular activity on ovarian biopsy (6), until a clear mechanism of hormone resistance is characterized. Although this high prevalence of follicular activity in POF is encouraging, we will have to await advances in the culture offollicles and ovum maturation in vitro to establish the relationship between these ultrasound findings and the presence of viable ova. Only time, therefore, will translate these results into practical advice for the prospects of fertility. 1.1 He 1 ~ S ~ Ii c '00.9 E c 'E g 0.8 III 0.7 ovaries not visualised * --- follicles < 4mm follicles > 4 mm reference group Figure 2 Bone mineral density measurements (mean and 95% confidence intervals for the mean) in women with POF divided into subgroups depending on whether no ovaries were identified on ultrasound (n = 22), follicles were present measuring <4 mm (n = 30) or >4 mm (n = 29) compared with the reference group (n = 57, P < , analysis of variance. *Significantly different from reference group by Duncan's procedure for multiple comparisons. **Significantly different from reference group and follicles >4 mm group. Fertility and Sterility

Estrogen status defined by ultrasound, using the "in vivo bioassay" of uterine size and endometrial thickness, was more informative than serum E2 concentrations as a predictor of the risk of

5 Estrogen status defined by ultrasound, using the "in vivo bioassay" of uterine size and endometrial thickness, was more informative than serum E2 concentrations as a predictor of the risk of osteoporosis. Thus, bone mineral density correlated with uterine cross-sectional are but not serum E2 measurements, possibly both reflecting the duration rather than the degree of estrogen deficiency (10). We have abandoned the P withdrawal test, which others also have found unhelpful in POF (3), in favor of the more informative ultrasound. Two subgroups of patients with POF in this series are worthy of further mention, those with primary amenorrhea and those with positive autoimmunity. Although patients presenting with primary amenorrhea had lower bone mineral density and smaller uterine size compared with those presenting with secondary amenorrhea, the two groups had a similar degree of autoimmunity and ovarian ultrasound findings. We conclude that patients with primary amenorrhea experience a longer duration of estrogen deficiency compared with women with secondary amenorrhea but that no difference in etiology has been defined by ultrasound. Patients with positive autoimmunity had similar ultrasound findings to those with a negative autoimmune screen. It may be that patients with a negative conventional autoimmune screen in fact harbor occult evidence for autoimmunity, such as autoantibodies directed perhaps against the FSH receptor (13) or oocyte (14), raised circulating immune complexes, or reduced natural killer cell activity (15) that were not sought in this study. Clearly, routine antiovarian antibodies are of little value in the investigation ofpof, as only one patient was positive for this test. This one patient also had Addison's disease and, in this context, antiovarian antibodies may be a useful predictor ofpof (4). The prevalence of positive autoimmunity is not only method dependent, as ovarian antibodies have been detected in 47% of 45 patients tested in one series using an ELISA (14), but also case dependent as the percentage of subjects with existing autoimmune disease varies from 16% to 60% in some series (13-15) compared with 10% in the present series. In conclusion, the majority of women with POF have follicles detectable by pelvic ultrasonography, which has replaced ovarian biopsy and P withdrawal tests in our investigation of such patients. The absence of ovarian activity on ultrasound may be a useful negative indicator of future prospects of ovarian functicm in POF. Patients presenting with primary amenorrhoea or with positive autoimmunity had a similar degree of ovarian activity on ultrasonography as the remainder of the group. Acknowledgments. We are grateful to Peter Ell, M.D., and the Department of Nuclear Medicine at the Middlesex Hospital, London, United Kingdom, for bone density measurements. REFERENCES 1. Coulam CB, Adamson SC, Annegers JF. Incidence of pre mature ovarian failure. Obstet GynecoI1986;67: Coulam CB, Ryan RJ. Prevalence of circulating antibodies directed toward ovaries among women with premature ovarian failure. Am J Reprod Immunol Microbiol 1985;9: Rebar RW, Connolly HV. Clinical features of young women with hypergonadotrophic amenorrhea. Fertil Steril 1990; 53: Betterle C, Rossi A, Pria SD, Artifoni A, Pedini B, Gavasso S, et al. Premature ovarian failure: autoimmunity and natural history. Clin Endocrinol (Oxf) 1993;39: Hague WM, Tan SL, Adams J, Jacobs HS. Hypergonadotrophic amenorrhea-etiology and outcome in 93 young women. Int J Gynaecol Obstet 1987;25: Jones GS, de Moraes-Reuhsen M. A new syndrome of amenorrhoea in association with hypergonadotropism and apparently normal ovarian follicular apparatus. Am J Obstet Gynecol 1969; 104: Dewhurst CJ, De Koos EB, Ferreira HP. The resistant ovary syndrome. Br J Obstet Gynaecol 1975;82: Lim HT, Meinders AE, de Haan LD, Bronkhurst FB. Anovulation presumably due to the gonadotrophin-resistant ovary syndrome. Eur J Obstet Gynaecol Reprod BioI 1984; 16: Freidman CI, Barrows H, Kim MH. Hypergonadotrophic hypogonadism. Am J Obstet GynecoI1983;145: Davies MC, Hall ML, Jacobs HS. Bone mineral loss in young women with amenorrhoea. Br Med J 1990;301: Sotsiou F, Bottazzo GF, Doniach D. Immunofluorescence studies on autoantibodies to steroid producing cells and to germline cells in endocrine disease and infertility. Clin Exp Immunol 1979; 39: Mehta AE, Matwijiw I, Lyons EA, Faiman C. Noninvasive diagnosis of resistant ovary syndrome by ultrasonography. Fertil Steril 1992;57: van Weissenbruch MM, Hoek A, van Vliet-Bleeker I, Schoemaker J, Drexhage H. Evidence for existence ofimmunoglobulins that block ovarian granulosa cell growth in vitro. A putative role in resistant ovary syndrome. J Clin Endocrinol Metab 1991;73: Luborsky JL, Visintin I, Boyers S, Asari T, Caldwell B, De Cherney A. Ovarian antibodies detected by immobilized antigen immunoassay in patients with premature ovarian failure. J Clin Endocrinol Metab 1990;70: Pekonen F, Siegberg R, Makinen T, Miettinen A, Yli-Korkala O. Immunological disturbances in patients with premature ovarian failure. Clin Endocrinol (Oxf) 1986;25:1-6. Conway et al. Characterization of idiopathic POF 341

Premature Menopause : Diagnosis and Management

Guideline Number 3 : August 2010 Premature Menopause : Diagnosis and Management Introduction : Premature menopause is a serious condition that affects young women and remains an enigma. The challenges